Receptor determinants of zoonotic transmission of …these pathogenic New World arenaviruses to humans. Our data further establish a critical biologic role for TfR1 in New World arenavirus

Receptor determinants of zoonotic transmissionof New World hemorrhagic fever arenavirusesSheli R. Radoshitzky*, Jens H. Kuhn*†, Christina F. Spiropoulou‡, Cesar G. Albarino‡, Dan P. Nguyen§,Jorge Salazar-Bravo¶, Tatyana Dorfman*, Amy S. Lee*, Enxiu Wang�, Susan R. Ross�, Hyeryun Choe§,and Michael Farzan*,**

*Department of Microbiology and Molecular Genetics and New England Primate Research Center, Harvard Medical School, Southborough, MA 01772;†Department of Biology, Chemistry, and Pharmacy, Freie Universitat Berlin, 14195 Berlin, Germany; ‡Special Pathogens Branch, Centers for DiseaseControl and Prevention, Atlanta, GA 30333; §Department of Pediatrics, Children’s Hospital, Harvard Medical School, Boston, MA 02115;¶Department of Biological Sciences, Center for Epidemiology and Zoonoses, Texas Tech University, Lubbock, TX 79409; and�Department of Microbiology and Abramson Family Cancer Center, University of Pennsylvania, Philadelphia, PA 19104

Edited by Peter Palese, Mount Sinai School of Medicine, New York, NY, and approved December 27, 2007 (received for review September 28, 2007)

Transferrin receptor 1 (TfR1) is a cellular receptor for the NewWorld hemorrhagic fever arenaviruses Machupo (MACV), Junın(JUNV), and Guanarito (GTOV). Each of these viruses is specificallyadapted to a distinct rodent host species, but all cause humandisease. Here we compare the ability of these viruses to use variousmammalian transferrin receptor 1 (TfR1) orthologs, including thoseof the South American rodents that serve as reservoirs for MACV,JUNV, and GTOV (Calomys callosus, Calomys musculinus, and Zy-godontomys brevicauda, respectively). Retroviruses pseudotypedwith MACV and JUNV but not GTOV glycoproteins (GPs) efficientlyused C. callosus TfR1, whereas only JUNV GP could use C. muscu-linus TfR1. All three viruses efficiently used Z. brevicauda TfR1. TfR1orthologs from related rodents, including house mouse (Musmusculus) and rat (Rattus norvegicus), did not support entry ofthese viruses. In contrast, these viruses efficiently used human anddomestic cat TfR1 orthologs. We further show that a local regionof the human TfR1 apical domain, including tyrosine 211, deter-mined the efficiency with which MACV, JUNV, and GTOV usedvarious TfR1 orthologs. Our data show that these New Worldarenaviruses are specifically adapted to the TfR1 orthologs of theirrespective rodent hosts and identify key commonalities betweenthese orthologs and human TfR1 necessary for efficient transmis-sion of these viruses to humans.

Calomys � Junın virus � Machupo virus � transferrin receptor 1

A renaviruses are enveloped, single-stranded, bisegmentedRNA viruses with ambisense genomes (1). The family Arena-

viridae consists of a single genus (Arenavirus) composed of at least24 viruses (2, 3). Based on their antigenic properties, arenaviruseshave been classified into two major groups: the Old World arena-viruses, which include lymphocytic choriomeningitis virus (LCMV)and Lassa virus (LASV), and the New World arenaviruses, whichare further divided into clades A, B, and C. The South Americanviruses Machupo (MACV), Junın (JUNV), Guanarito (GTOV),and Sabia (SABV) belong to clade B and cause Bolivian, Argen-tinian, Venezuelan, and Brazilian hemorrhagic fevers, respectively.MACV, JUNV, and GTOV are classified as National Institute ofAllergy and Infectious Disease Category A Priority Pathogens,Select Agents, and Class 4 Biosafety Pathogens, in part due to theirhigh lethality (2, 4).

Rodents of the Muridae family are the natural hosts of mostarenaviruses, and the geographic distribution of each arenavirus isdetermined by the range of its corresponding host. New Worldarenaviruses are found in the murid subfamily Sigmodontinae inspecialized ecologic niches in South and North America (5, 6).Calomys callosus (large vesper mouse), Calomys musculinus (dry-lands vesper mouse), and Zygodontomys brevicauda (cane mouse)are the principal hosts for MACV, JUNV, and GTOV, respectively.The host of SABV has not been identified (7–10). The phylogeneticdiversity of arenaviruses is likely the result of long-term coevolutionof the viruses and their corresponding hosts (11, 12).

Entry of arenaviruses into their target cells is facilitated by thetwo noncovalently linked surface glycoproteins GP1 and GP2. Bothproteins are synthesized in infected cells as a single glycoproteinprecursor (GPC), which is proteolytically processed to these twomature subunits (13–16). As with other class I fusion proteins, theGP1 subunit associates with a cellular receptor (17–20). The GP2subunit is a transmembrane protein that mediates fusion of the viraland cellular membranes after internalization of the virus intoacidified endosomes (21–24).

�-Dystroglycan is a cellular receptor for New World clade C andOld World arenaviruses (25–27). Four New World clade B arena-viruses (MACV, JUNV, GTOV, and SABV) use transferrin re-ceptor 1 (TfR1) as a cellular receptor (28). Several properties ofTfR1 support its critical role in arenaviral replication and disease.It is rapidly and constitutively internalized by clathrin-mediatedendocytosis into an acidic compartment, consistent with the path-way and pH dependence of arenavirus cell entry (23, 29). It isexpressed ubiquitously and is expressed at high levels on activatedor rapidly dividing cells, including macrophages and activatedlymphocytes, which are major targets of arenaviruses in vivo(30–33). TfR1 is also highly expressed on endothelial cells (34–36),thought to be central to the pathogenesis of hemorrhagic fevers (35,37). TfR1 up-regulation on immune cells activated in response toinfection may accelerate viral replication in these cells and may inpart explain the higher lethality of New World hemorrhagic feverscompared with Lassa fever. Several studies also indicate thatadditional or alternative receptors for clade B viruses exist (38–41).For example, the nonpathogenic clade B viruses Amapari andTacaribe use a receptor on human cells distinct from TfR1 and�-dystroglycan (38, 41).

Here we describe the ability of MACV, JUNV, and GTOV GPto mediate entry into cells expressing a range of TfR1 orthologs. Weobserved that JUNV and MACV but not GTOV efficiently usedthe TfR1 ortholog of the MACV host species, C. callosus, whereasonly JUNV used the TfR1 ortholog of its host species, C. muscu-linus. Although GTOV could not use either Calomys species TfR1ortholog, it efficiently used that of its principal reservoir, Z.brevicauda. House mouse (Mus musculus), rat (Rattus norvegicus),and dog TfR1 orthologs were inefficient receptors for all threeviruses, whereas cat and human TfR1 supported their efficient

Author contributions: S. R. Radoshitzky and M.F. designed research; S. R. Radoshitzky,J.H.K., C.F.S., C.G.A., D.P.N., T.D., A.S.L., E.W., and H.C. performed research; J.S.-B., E.W., andS. R. Ross contributed new reagents/analytic tools; S. R. Radoshitzky and M.F. analyzed data;and S. R. Radoshitzky and M.F. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The sequences reported in this paper have been deposited in the GenBankdatabase (accession nos. EU164540, EU164541, and EU340259).

**To whom correspondence should be addressed. E-mail: [email protected].

© 2008 by The National Academy of Sciences of the USA

2664–2669 � PNAS � February 19, 2008 � vol. 105 � no. 7 www.pnas.org�cgi�doi�10.1073�pnas.0709254105

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020

entry. Six residues from the apical domain of human TfR1 con-verted house mouse TfR1 to an efficient receptor for each of thesearenaviruses. These studies localized the GP1-binding region to theC terminus of �-helix 2 and to a prominent loop between � strands1 and 2 in the apical domain of TfR1. This latter loop includestyrosine 211, which is common to human, cat, C. callosus, C.musculinus, and Z. brevicauda TfR1 orthologs but absent from ratand house mouse orthologs. We conclude that similarities in thisregion between human TfR1 and the TfR1 orthologs of the MACV,JUNV, and GTOV host species facilitate efficient transmission ofthese pathogenic New World arenaviruses to humans. Our datafurther establish a critical biologic role for TfR1 in New Worldarenavirus replication and South American hemorrhagic fevers.

ResultsUtilization of TfR1 Orthologs by New World Hemorrhagic Fever Arena-viruses. To obtain insight into zoonotic transmission of New Worldarenaviruses and to help identify TfR1 determinants critical to its

role as an arenaviral receptor, we compared human TfR1 (hTfR1)to TfR1 orthologs from M. musculus (house mouse; mTfR1), R.norvegicus (rat; rTfR1), Canis familiaris (dog; cTfR1), Felis domes-ticus (cat; fTfR1), C. callosus (large vesper mouse; ccTfR1), C.musculinus (drylands vesper mouse; cmTfR1), and Z. brevicauda(cane mouse; zbTfR1). C. callosus, C. musculinus, and Z. brevicaudaserve as reservoirs for MACV, JUNV, and GTOV, respectively (5).Chinese hamster ovary (CHO) cells, which are refractory to trans-duction by Moloney murine leukemia virus (MLV) pseudotypedwith the glycoproteins of MACV, JUNV, or GTOV (28, 41), weretransfected with plasmids expressing these TfR1 orthologs taggedin their C-terminal ectodomains. Transfected CHO cells then werecharacterized for TfR1 expression by flow cytometry. In parallel, achimeric protein in which residues 79–258 of MACV GP1 werefused to the Fc region of human IgG1 (MACV GP1�-Fc) wasassayed for its ability to bind transfected cells. This truncationvariant has been previously shown to bind human TfR1 with higheraffinity than full-length GP1-Fc (28). Finally, transfected cells weretransduced with MLV expressing GFP and pseudotyped with theGP of MACV, JUNV, or GTOV (MACV-MLV, JUNV-MLV, orGTOV-MLV, respectively), and cell entry was measured as GFPfluorescence. When roughly equivalent cell surface expression ofeach TfR1 ortholog was observed (Fig. 1A), MACV GP1�-Fcefficiently bound hTfR1, fTfR1, and ccTfR1 but not mTfR1, rTfR1,cTfR1, or cmTfR1 (Fig. 1B). The ability of MACV GP1�-Fc to bindcells transfected with TfR1 orthologs correlated with the ability ofMACV-MLV to enter these cells (Fig. 1C). Similarly, neitherJUNV-MLV nor GTOV-MLV could enter cells expressing mTfR1,rTfR1, or cTfR1. Strikingly, cells expressing fTfR1 were moreefficiently transduced by JUNV-MLV and GTOV-MLV thanhTfR1-expressing cells. Expectedly, MACV-MLV and JUNV-MLV each efficiently entered cells expressing the TfR1 orthologsof their respective host species, ccTfR1 and cmTfR1. However,MACV GP could not use cmTfR1. Also, the GTOV GP did notmediate entry into cells expressing either ccTfR1 or cmTfR1 (Fig.1 C and E) (5). However, it did mediate efficient entry into cellsexpressing the TfR1 of the GTOV host species Z. brevicauda, as didMACV and JUNV GP molecules (Fig. 1E). These data demon-strate that MACV, JUNV, and GTOV are specifically adapted tothe TfR1 orthologs of their respective hosts. The unexpectedobservation that feline TfR1 is a highly efficient receptor for eachof these viruses raises the possibility that a felid, perhaps a predatorof these rodent hosts, may serve as a transmission intermediate forone or more New World hemorrhagic fever arenaviruses.

Localization of the Arenaviral GP1-Binding Determinants on HumanTfR1. Next, we characterized a series of chimeras of hTfR1 andmTfR1 (represented in Fig. 2A) for their ability to bind MACVGP1�-Fc and support MACV, JUNV, and GTOV GP-mediatedentry. Chimeras 1–6 have been previously described (42). Allreceptor chimeras were expressed on the cell surface to similarlevels as determined by flow cytometry (Fig. 2B). As shown in Fig.2C, MACV GP1�-Fc efficiently bound human TfR1 as well as aTfR1 chimera consisting of the hTfR1 apical domain and themTfR1 protease-like and helical domains (chimera 7), thus local-izing the GP1-binding region to the hTfR1 apical domain. Com-parison of chimera 5, which did not bind GP1�-Fc, with chimera 6,which did, identifies a binding determinant between residues 325and 366 of the TfR1 apical domain. Comparison of chimeras 7 and8, which differ only by the presence of mTfR1 amino acids betweenresidues 208 and 212, identifies a second potential binding deter-minant. Unlike chimera 7, chimera 8 did not associate with GP1�-Fc, suggesting that these residues in hTfR1 bind MACV GP1. Theability of chimeras 9 and 10 to bind MACV GP1�-Fc efficiently isconsistent with a role for these two GP1-binding determinants (Fig.2C). Chimera 9 contains human sequences that include bothregions (188–229 include 208–212, and 325–383 include 325–366)in a largely murine receptor. Chimera 10 also contains these

Fig. 1. New World arenavirus entry mediated by TfR1 orthologs. (A) CHO cellsweretransfectedwithplasmidsencodingTfR1orthologs fromhuman(hTfR1),M.musculus (mTfR1), R. norvegicus (rTfR1), C. familiaris (cTfR1), F. domesticus(fTfR1), C. callosus (ccTfR1), or C. musculinus (cmTfR1) tissues. Cell surface expres-sion was determined by flow cytometry using an anti-FLAG antibody recognizinga tag present at the C terminus of each of the TfR1 variants. (B) In parallel,transfected cells were incubated with 200 nM of the MACV GP1 truncationvariant, GP1�, fused to the Fc domain of human IgG1 (GP1�-Fc), and TfR1association was determined by flow cytometry. Mean fluorescence values werenormalized to those of hTfR1. Error bars indicate the standard deviation of fiveexperiments. (C) An aliquot of the cells used in A and B were transduced withMLVs expressing enhanced GFP and pseudotyped with the glycoproteins ofMACV, JUNV, or GTOV. Forty-eight hours after transduction, cell entry wasmeasured by flow cytometry. Mean fluorescence values were normalized tothose of hTfR1-expressing cells. Error bars indicate the standard deviation of fiveexperiments. (D) An experiment similar to that in A, except that CHO cells weretransfected with plasmid encoding the Z. brevicauda TfR1 ortholog (zbTfR1), aswell as those encoding hTfR1, mTfR1, ccTfR1, and cmTfR1. (E) An experimentsimilar to that in C in which cell entry efficiency was measured by using an aliquotof the cells used in D.

Radoshitzky et al. PNAS � February 19, 2008 � vol. 105 � no. 7 � 2665

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020

regions. TfR1 chimeras that bound GP1�-Fc efficiently (6, 7, 9, and10) also supported efficient entry mediated by MACV, JUNV, orGTOV GP (Fig. 2D). These data indicate that arenaviral GP-mediated entry depends on at least two determinants in the hTfR1apical domain that are separated in the TfR1 primary sequence butadjacent in the tertiary structure (43). Consistent with a role for theTfR1 apical domain in arenaviral entry, an anti-human TfR1antibody previously shown to inhibit replication of infectiousMACV, JUNV, GTOV, and SABV (28) recognized chimera 7 butnot mTfR1 or chimera 8 (data not shown). Thus, this antibody, likeMACV GP1�-Fc, recognizes the human TfR1 apical domain,probably including hTfR1 residues 208–212. Also consistent with acentral role for the TfR1 apical domain in arenaviral entry, neithersoluble transferrin (28) nor overexpressed human hemochromato-sis protein HFE (not shown) altered the efficiency of MACV entry.These TfR1-binding proteins associate with TfR1 helical andprotease-like domains and do not contact the apical domain(44, 45).

Tyrosine 211 in the Human TfR1 Apical Domain Is a Critical Determi-nant of MACV, JUNV, and GTOV Cell Entry. Comparison of residues208–212 and 325–366 among the TfR1 orthologs characterized inFig. 1 suggested that tyrosine 211 might be a critical determinant ofMACV, JUNV, and GTOV entry and of variation in host suscep-tibility to these viruses. Most TfR1 orthologs with a tyrosine at thisposition supported entry of at least one of these arenaviruses,whereas orthologs lacking this tyrosine did not function as arena-viral receptors (critical differences among TfR1 orthologs arerepresented in summary in Fig. 3). Alteration of tyrosine 211 ofhTfR1 to aspartic acid or alanine prevented surface expression ofhTfR1. However, one hTfR1 variant, in which tyrosine 211 wasaltered to threonine (Y211T), expressed efficiently (Fig. 4A). Wealso generated human TfR1 variants in which asparagine 292 andasparagine 348 were altered to their mTfR1 counterparts, glutamicacid (N292E) and lysine (N348K), respectively. Asparagine 348 isadjacent to tyrosine 211 in the tertiary structure of hTfR1 (Fig. 3C).MACV GP1�-Fc efficiently bound hTfR1 and the N292E hTfR1variant. In contrast, hTfR1 variants N348K and Y211T did notsupport MACV GP1�-Fc binding (Fig. 4B). Again, these observa-

tions are consistent with the ability of these receptor variants tomediate entry of MACV-MLV, JUNV-MLV, and GTOV-MLV(Fig. 4C). Together, these data suggest a critical role for residues211 and 348 of human TfR1 in mediating entry of New Worldhemorrhagic fever arenaviruses. Fig. 3C shows the crystal structureof human TfR1 apical domain (43) in which residues 211 and 348,proximal in the hTfR1 tertiary structure, are indicated.

Introduction of Six hTfR1 Residues Convert mTfR1 to an Efficient NewWorld Arenavirus Receptor. Tyrosine 211 is located within an ex-posed loop of hTfR1 between �-strands 1 and 2. We introduced thisfive-residue loop (RLVYL) of hTfR1 into mTfR1, replacing itsfour-residue loop (NLDP). We also constructed a second mTfR1variant in which lysine 348 was altered to asparagine (K348N). Athird mTfR1 variant (RLVYL�K348N) includes both of thesealterations. Although all variants expressed as efficiently as hTfR1and mTfR1 (Fig. 5A), only mTfR1 variants with the RLVYL loopbound MACV GP1�-Fc (Fig. 5B). GP1�-Fc associated mostefficiently with RLVYL�K348N. Similarly, MACV-MLV andGTOV-MLV could enter cells expressing the RLVYL mTfR1variant and, more efficiently, the RLVYL�K348N mTfR1 variant(Fig. 5C). The RLVYL variant mediated JUNV-MLV entry asefficiently as hTfR1, and the K348N change had no additionaleffect. These data show that hTfR1 residues 208–212, includingtyrosine 211, are critical for MACV, JUNV, and GTOV entry andthat MACV and GTOV entry is partially disrupted by lysine 348,which is present in mTfR1. JUNV may better tolerate this lysinebecause the TfR1 of its host species, C. musculinus, also has a lysineat position 348 (Fig. 3D).

Effect of Potential N-glycosylation of C. callosus and C. musculinusTfR1 Asparagine 205 on MACV-MLV, JUNV-MLV, and GTOV Entry. C.callosus, C. musculinus, and Z. brevicauda TfR1 possess apotential N-glycosylation site proximal to tyrosine 211, whereashuman TfR1 lacks such a site (Fig. 3D). We constructed ccTfR1and cmTfR1 variants lacking this glycosylation motif at aspar-agine 205 (ccTfR1 N205A and cmTfR1 N205A) and assayedtheir ability to mediate entry of MACV-, JUNV-, and GTOV-MLV (Fig. 6 A and B). Removal of the glycosylation motif from

Fig. 2. MACV, JUNV, and GTOV entry medi-ated by chimeras of human and mouse TfR1.(A) A representation of the TfR1 structuraldomains and human/mouse TfR1 chimeras. Inthe top bar, the protease-like, apical, and he-lical domains of human TfR1 are indicated asblue, red, and cyan, respectively. The N-terminal cytoplasmic domain and transmem-brane domain are shown in white. Individualmouse/human chimeras are represented ingray, indicating murine sequence, and inwhite, indicating human sequence. A plus signto the right of each chimera indicates efficientMACV GP1�-Fc association and MACV, JUNV,and GTOV GP-mediated entry, shown in C andD. (B) CHO cells were transfected with plasmidsencoding hTfR1, mTfR1, and chimeras of thesereceptors. Cell surface expression was analyzedas in Fig. 1 A. Mean fluorescence values werenormalized to hTfR1. Error bars indicate thestandard deviation of three experiments. (C) Inparallel, cell surface binding of MACV GP1�-Fcwas determined by flow cytometry, as in Fig.1B. Mean fluorescence values were normalizedto those of hTfR1-expressing cells. Error barsindicate the standard deviation of three exper-iments. (D) An aliquot of the cells used in B andC was transduced with MACV, JUNV, or GTOVpseudoviruses and analyzed as in Fig. 1C. Mean fluorescence values were normalized to hTfR1. Error bars indicate the standard deviation of threeexperiments.

2666 � www.pnas.org�cgi�doi�10.1073�pnas.0709254105 Radoshitzky et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020

ccTfR1 had a modest effect on MACV entry (�30% increase).However, this variant substantially increased JUNV GP-mediated entry (�2-fold over wild-type ccTfR1) and, unlikewild-type ccTfR1, was an efficient receptor for GTOV-MLV.Removal of the glycosylation motif from cmTfR1 increased theentry efficiency of each virus but to a lesser extent than observedwith ccTfR1. These data suggest that the efficiency of hTfR1 asa receptor for MACV, JUNV, and GTOV may in part be due tothe absence of this glycosylation site. They also underscore thegreater efficiency with which these viruses use the TfR1 or-thologs of their natural hosts, despite apparent interference fromglycosylation at asparagine 205. The analogous residue of hTfR1,aspartic acid 204, is shown in Fig. 3C.

DiscussionTfR1 is a cellular receptor for New World hemorrhagic feverarenaviruses (28) and is likely to play a critical role in their zoonotictransmission and the pathogenesis of arenaviral hemorrhagic fe-vers. Here we identify the GP1-binding site on human TfR1.Specifically, residues 208–212, which are localized to a TfR1apical-domain loop between �-strands 1 and 2, and residue 348,which is adjacent to this loop (Fig. 3C), participate in MACV GP1binding and in MACV, JUNV, and GTOV GP-mediated entry intotarget cells. By introducing these residues into house mouse (M.musculus) TfR1, we converted it to an efficient receptor for NewWorld hemorrhagic fever arenaviruses. These results are consistentwith the presence of tyrosine 211 in TfR1 orthologs from the

MACV, JUNV, and GTOV rodent hosts, C. callosus, C. musculi-nus, and Z. brevicauda, respectively (Fig. 3D). Tyrosine 211 ispresent in cat TfR1, a very efficient receptor for all three NewWorld hemorrhagic fever arenaviruses. TfR1 from the rhesusmacaque, a non-human primate model for MACV and JUNVinfections (46, 47), also has a tyrosine at position 211. In contrast,position 211 is an aspartic acid in house mouse and rat TfR1orthologs, preventing efficient arenavirus entry. Furthermore,hamster TfR1 also possesses an aspartic acid at this position,consistent with the resistance of hamster cell lines, including CHOand BHK cells, to New World arenavirus GP-mediated transduc-tion (28, 41). Thus, the presence of tyrosine 211 in the humanreceptor is likely to be critical for zoonotic transmission of hemor-rhagic fever arenaviruses.

Although tyrosine 211 is necessary for New World arenaviralinfection, it is not sufficient. The dog TfR1 ortholog, which ex-presses this tyrosine, does not function as an efficient receptor forMACV, JUNV, or GTOV. Neither MACV nor GTOV uses theTfR1 ortholog of the JUNV host species, C. musculinus, which alsohas this tyrosine. Residues in the local vicinity of tyrosine 211 likelyaccount for these differences. For example, dog TfR1 has two acidicresidues at positions 206 and 208 (Fig. 3D) that are not present inreceptors that mediate efficient arenavirus entry. Similarly, C.musculinus TfR1 includes a lysine at residue 348, which we haveshown to interfere with MACV and GTOV entry but which doesnot affect entry of JUNV (Fig. 5C).

Fig. 3. TfR1 determinants of zoonotic transmission of New World hemorrhagicfever arenaviruses. (A) The structure of the human TfR1 dimer is shown orientedwith the cellular membrane at the bottom. Protease-like, apical, and helicaldomains are indicated as blue, red, and cyan, respectively, on one monomer. Theother monomer is shown in white. A loop composed of residues 208–212, criticalfor New World arenaviral entry, is indicated in green. (B) As in A, except thathuman TfR1 dimer is rotated 150° about the twofold axis of the dimer. (C) As inB, except that the apical domain is enlarged. Aspartic acid 204, corresponding toa potential glycosylation site in the Calomys species, is indicated in blue. Aspar-agine 348, necessary together with residues 208–212 to convert mTfR1 to anefficient MACV and GTOV receptor, is shown in yellow. Tyrosine 211, within the208–212 loop, is also indicated. (D) An alignment of amino acid sequences fromtwo proximal regions of the indicated TfR1 orthologs. Human TfR1 residues thatconvert mouse TfR1 to an efficient MACV, JUNV, and GTOV receptor are under-lined. Tyrosine 211 and lysine 348 are shown in green and red, respectively.Potential glycosylation sites present in Calomys species and Z. brevicauda TfR1orthologs are indicated in blue. Macaca mulatta (rhesus macaque) and Cricetulusgriseus (Chinese hamster) TfR1 regions are shown with those of the TfR1 or-thologs characterized here. Rhesus macaques can be used as a model for MACVand JUNV infection (46, 47, 52). Hamster CHO and BHK cell lines are refractory toentry mediated by MACV, JUNV, and GTOV GP (28, 41).

Fig. 4. The arenavirus GP1-binding site of human TfR1. (A) CHO cells weretransfected with plasmids encoding hTfR1, mTfR1, or human TfR1 variantsN292E, N348K, or Y211T. Cell surface expression was analyzed as in Fig. 1. MeanfluorescencevalueswerenormalizedtothoseofhTfR1-expressingcells.Errorbarsindicate the standard deviation of three experiments. (B) In parallel, cell surfacebinding of MACV GP1�-Fc was determined by flow cytometry. Mean fluores-cence values were normalized to those of hTfR1-expressing cells. Error barsindicate the standard deviation of three to five experiments. (C) An aliquot of thecells used in A and B was transduced with MACV, JUNV, or GTOV pseudovirusesand analyzed as in Fig. 1C. Mean fluorescence values were normalized to hTfR1.Error bars indicate the standard deviation of three experiments.

Radoshitzky et al. PNAS � February 19, 2008 � vol. 105 � no. 7 � 2667

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020

These arenaviruses are specific for the TfR1 orthologs of theirrespective host species but, paradoxically, also use the less relatedcat and human orthologs. For example, MACV uses the C. callosusTfR1 efficiently, but it cannot use the closely related C. musculinusTfR1. GTOV cannot use TfR1 from MACV or JUNV host species.Nonetheless, all three viruses use human TfR1 efficiently, despitethe lack of obvious opportunity to adapt to the human receptor.Our data show that a potential glycosylation site adjacent to theGP-binding site of the C. callosus and C. musculinus TfR1 interfereswith entry of MACV, JUNV, and GTOV (Fig. 6B). Human and catTfR1 lack this glycosylation motif which, combined with sequencesimilarities to Calomys and Zygodontomys TfR1 orthologs aroundtyrosine 211, may account for the relatively efficient use of thesereceptors.

Our data also have implications for the future study of arenaviralhemorrhagic fevers and for efforts to prevent or treat these fevers.First, use of a common binding region by MACV, JUNV, andGTOV suggests that a single antibody or small molecule, includingmimetics of the GP-binding site, may be useful in controllingreplication of all three viruses. Second, our observations may beuseful in identifying species that may serve as additional reservoirsor intermediates in the transmission of New World arenaviruses.Third, our observations suggest an approach for generating an adultmurine model of arenaviral hemorrhagic fever, namely by modify-ing the murine Tfr1 gene to include the six residues identified here.

Finally, our data underscore the biologic role of TfR1 in NewWorld arenavirus replication. We have shown that MACV, JUNV,and GTOV GPs use the TfR1 orthologs of their respective hostsspecifically and with high efficiency. Similarities between the GP-binding sites of these TfR1 orthologs with hTfR1 provide anexplanation for the efficient zoonotic transmission of these patho-genic arenaviruses. The absence of a house mouse or rat model ofSouth American hemorrhagic fevers (48–50) is consistent with theinefficiency of mouse and rat TfR1 orthologs as receptors for

MACV, JUNV, and GTOV. Collectively, these data indicate thatTfR1 is central to the replication of these viruses in humans andtheir natural hosts.

Materials and MethodsCells and Plasmids. Human embryonic kidney 293T cells [American Type CultureCollection (ATCC) no. CRL-11268] were maintained in Dulbecco modified Eagle’smedium, and CHO (ATCC no.CCL-61) epithelial cells in Ham F12 medium. Both celllines were supplemented with 10% fetal bovine serum (Sigma), 100 units/mlpenicillin, and 100 �g/ml streptomycin (Cellgro). Plasmids encoding F. domesticus(cat) and C. familiaris (dog) TfR1 (fTfR1 and cTfR1, respectively) were generouslyprovided by Colin Parrish (Cornell University, Ithaca, NY). R. norvegicus TfR1(rTfR1) was cloned from a rat liver cDNA library (BioChain) by using the followingprimers: 5�-AATAACTACTTCGAAGCCACCATGGATCAAGCCAGATCAGCATT-CTC-3� and 5�-AATAACTACCTCGAGTTAAAACTCATTGTCAATATTCCAAATGTC-ACCAG-3�.

C. callosus (large vesper mouse), C. musculinus (drylands vesper mouse),and Z. brevicauda Tfr1 genes (ccTfR1, cmTfR1, and zbTfR1, respectively)were cloned from frozen liver tissues (GenBank accession nos. EU164540,EU164541, and EU340259, respectively). RNA was isolated from tissues byusing an RNAqueous kit (Ambion), and cDNA was generated by using theSuperScript III First-Strand Synthesis system for RT-PCR (Invitrogen). Thefollowing primers were used to amplify the cmTfR1, ccTfR1, and zbTfR1genes: 5�-ACAACTATGATGGATCAAGCCAGATCAGCA-3� and 5�-ACAAC-TACATTTAAAACTCATTGTCAATATTCCAAATGTC-3�. Coding regions ofhuman, M. musculus (house mouse) (28), F. domesticus, C. familiaris, R.norvegicus, C. callosus, C. musculinus, and Z. brevicauda TfR1 were clonedinto the pcDNA 3.1(�) expression vector (Invitrogen) with a C-terminalFLAG tag. Chimeras of M. musculus and human TfR1 (mTfR1 and hTfR1,respectively) were generated using gene splicing by overlap extension (51).All chimeras were cloned into the pcDNA 3.1(�) expression vector (Invitro-gen) with a C-terminal FLAG tag. Site-directed mutagenesis of mTfR1 andhTfR1 was performed by using Expand Long-Range dNTPack (Roche).Plasmids encoding the MACV Carvallo strain GP1� deletion variant (resi-dues 79 –248) fused to the Fc region of human IgG1 (GP1�-Fc), as well as

Fig. 5. Conversion of house mouse TfR1 to an efficient arenavirus receptor.(A) CHO cells were transfected with plasmids encoding hTfR1, mTfR1, andmTfR1 mutants (RLVYL, K348N, and RLVYL�K348N). Cell surface expressionwas analyzed as in Fig. 1. Mean fluorescence values were normalized to hTfR1.Error bars indicate the standard deviation of three experiments. (B) In parallel,cell surface binding of MACV GP1�-Fc was determined by flow cytometry.Mean fluorescence values were normalized to hTfR1. Error bars indicate thestandard deviation of three to five experiments. (C) An aliquot of the cells usedin A and B was transduced with MACV, JUNV, or GTOV pseudoviruses andanalyzed as in Fig. 1C. Mean fluorescence values were normalized to hTfR1.Error bars indicate the standard deviation of three experiments.

Fig. 6. Effect of C. callosus and C. musculinus asparagine 205 glycosylation onMACV, JUNV, and GTOV entry. (A) CHO cells were transfected with plasmidsencoding hTfR1, C. callosus TfR1 (ccTfR1), C. musculinus TfR1 (cmTfR1), or N205Aglycosylation mutants (ccTfR1 N205A or cmTfR1 N205A). Cell surface expressionwas analyzed as in Fig. 1. Mean fluorescence values were normalized to hTfR1.Error bars indicate the standard deviation of three experiments. (B) An aliquot ofthe cells used in A was transduced with MACV, JUNV, or GTOV pseudoviruses andanalyzed as in Fig. 1C. Mean fluorescence values were normalized to hTfR1. Errorbars indicate the standard deviation of three experiments.

2668 � www.pnas.org�cgi�doi�10.1073�pnas.0709254105 Radoshitzky et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020

MACV, JUNV (MC2), and GTOV (INH-95551) GPC, have been describedpreviously (28).

Binding Assays. To generate the MACV GP1�-Fc fusion protein, 293T cells weretransfected with the appropriate plasmid according to the calcium phosphatemethod. The protein was purified from SFM II (Invitrogen) with protein A-Sepharose beads, eluted with 3 M MgCl2, dialyzed in PBS, and concentrated.Association of MACV GP1�-Fc with TfR1 orthologs and variants was determinedby flow cytometry. CHO cells were transfected with the different TfR1-encodingplasmids or with vector alone with FuGENE HD transfection reagent (Roche).Twenty-four hours after transfection, cells were detached with trypsin andseeded in 6- or 24-well plates for subsequent flow cytometric experiments assay-ing cell surface expression, MACV GP1�-Fc association, and pseudovirus trans-duction. After 24 h, CHO cells expressing TfR1 orthologs and variants weredetached with cell dissociation buffer (GIBCO) and washed with 2% goat serumin PBS. Anti-FLAG M2 murine antibody (Sigma) or 200 nM MACV GP1�-Fc proteinwasaddedto5�105 cells,andthemixturewas incubatedat4°Cfor1h.Cellswerewashed once with PBS/2% goat serum and incubated for 45 min at 4°C withanti-mouse IgG-specific phycoerythrin conjugate (Sigma) or with anti-human IgG(Fc-specific) FITC conjugate (Sigma). Cells were again washed with PBS/2% goatserum, and cell surface binding was detected by flow cytometry, with 10,000events counted per sample. Background fluorescence was determined by mea-

suring cells treated only with anti-mouse IgG-specific PE conjugate or withanti-human IgG (Fc-specific) FITC conjugate. Each assay was performed in dupli-cate and repeated three to five times.

Pseudovirus Transduction. To generate retroviruses pseudotyped with arenaviralGP, 293T cells were transfected according to the calcium phosphate method withplasmids encoding MACV GPC, JUNV GPC, or GTOV GPC, together with thepQCXIX vector (BD Biosciences) expressing GFP, and plasmid encoding the MLVgag and pol genes by using equal concentrations of each plasmid. Cell superna-tants were harvested 48 h after transfection, cleared of cellular debris by centrif-ugation, filtered through a 0.45-�m pore size filter (Corning Glass), and stored at�80°C. Supernatants containing pseudoviruses were added to CHO cells trans-fected to express TfR1 orthologs and variants in 24-well plates. After 1.5 h ofincubation at 37°C, cells were washed once with PBS and replenished with freshmedia. At 48 h after infection, cells were imaged by fluorescence microscopy anddetached with trypsin for GFP fluorescence analysis by flow cytometry. Fluores-cence was normalized to mock-transfected cells transduced with each of thepseudoviruses.Eachassaywasperformed induplicatesandrepeatedthreetofivetimes.

ACKNOWLEDGMENTS. We thank Thomas Postler for comments and carefulmanuscripteditingandColinParishforhelpfuladviceandforplasmidsexpressingcat and dog TfR1.

1. Oldstone MB (2002) Arenaviruses I: The epidemiology molecular and cell biology ofarenaviruses. Curr Top Microbiol Immunol 262:V–XII.

2. Charrel RN, de Lamballerie X (2003) Arenaviruses other than Lassa virus. Antiviral Res57:89–100.

3. Lecompte E, Ter Meulen J, Emonet S, Daffis S, Charrel RN (2007) Genetic identificationof Kodoko virus, a novel arenavirus of the African pigmy mouse (Mus Nannomysminutoides) in West Africa. Virology 364:178–183.

4. Borio L, et al. (2002) Hemorrhagic fever viruses as biological weapons: medical andpublic health management. J Am Med Assoc 287:2391–2405.

5. Salazar-Bravo J, Ruedas LA, Yates TL (2002) Mammalian reservoirs of arenaviruses. CurrTop Microbiol Immunol 262:25–63.

6. Clegg JC (2002) Molecular phylogeny of the arenaviruses. Curr Top Microbiol Immunol262:1–24.

7. Johnson KM, Mackenzie RB, Webb PA, Kuns ML (1965) Chronic infection of rodents byMachupo virus. Science 150:1618–1619.

8. Johnson KM, Kuns ML, Mackenzie RB, Webb PA, Yunker CE (1966) Isolation of Ma-chupo virus from wild rodent Calomys callosus. Am J Trop Med Hyg 15:103–106.

9. Fulhorst CF, et al. (1999) Natural rodent host associations of Guanarito and piritalviruses (Family Arenaviridae) in central Venezuela. Am J Trop Med Hyg 61:325–330.

10. Mills JN, et al. (1991) Junin virus activity in rodents from endemic and nonendemic lociin central Argentina. Am J Trop Med Hyg 44:589–597.

11. Hugot JP, Gonzalez JP, Denys C (2001) Evolution of the Old World Arenaviridae andtheir rodent hosts: Generalized host-transfer or association by descent? Infect GenetEvol 1:13–20.

12. Bowen MD, Peters CJ, Nichol ST (1997) Phylogenetic analysis of the Arenaviridae:Patterns of virus evolution and evidence for cospeciation between arenaviruses andtheir rodent hosts. Mol Phylogenet Evol 8:301–316.

13. Buchmeier MJ, Southern PJ, Parekh BS, Wooddell MK, Oldstone MB (1987) Site-specificantibodies define a cleavage site conserved among arenavirus GP-C glycoproteins.J Virol 61:982–985.

14. Kunz S, Edelmann KH, de la Torre JC, Gorney R, Oldstone MB (2003) Mechanisms forlymphocytic choriomeningitis virus glycoprotein cleavage, transport, and incorpora-tion into virions. Virology 314:168–178.

15. Lenz O, ter Meulen J, Klenk HD, Seidah NG, Garten W (2001) The Lassa virus glycop-rotein precursor GP-C is proteolytically processed by subtilase SKI-1/S1P. Proc Natl AcadSci USA 98:12701–12705.

16. Beyer WR, Popplau D, Garten W, von Laer D, Lenz O (2003) Endoproteolytic processingof the lymphocytic choriomeningitis virus glycoprotein by the subtilase SKI-1/S1P.J Virol 77:2866–2872.

17. Hughson FM (1995) Structural characterization of viral fusion proteins. Curr Biol 5:265–274.18. Kielian M, Rey FA (2006) Virus membrane-fusion proteins: More than one way to make

a hairpin. Nat Rev Microbiol 4:67–76.19. Kunz S, Borrow P, Oldstone MB (2002) Receptor structure, binding, and cell entry of

arenaviruses. Curr Top Microbiol Immunol 262:111–137.20. Schibli DJ, Weissenhorn W (2004) Class I, class II viral fusion protein structures reveal

similar principles in membrane fusion. Mol Membr Biol 21:361–371.21. Gallaher WR, DiSimone C, Buchmeier MJ (2001) The viral transmembrane superfamily:

Possible divergence of Arenavirus and Filovirus glycoproteins from a common RNAvirus ancestor. BMC Microbiol 1:1.

22. York J, Agnihothram SS, Romanowski V, Nunberg JH (2005) Genetic analysis of heptad-repeat regions in the G2 fusion subunit of the Junin arenavirus envelope glycoprotein.Virology 343:267–274.

23. Castilla V, Mersich SE, Candurra NA, Damonte EB (1994) The entry of Junin virus intoVero cells. Arch Virol 136:363–374.

24. Eschli B, et al. (2006) Identification of an N-terminal trimeric coiled-coil core withinarenavirus glycoprotein 2 permits assignment to class I viral fusion proteins. J Virol80:5897–5907.

25. Cao W, et al. (1998) Identification of alpha-dystroglycan as a receptor for lymphocyticchoriomeningitis virus and Lassa fever virus. Science 282:2079–2081.

26. Kunz S, Rojek JM, Perez M, Spiropoulou CF, Oldstone MB (2005) Characterization of theinteraction of lassa fever virus with its cellular receptor alpha-dystroglycan. J Virol79:5979–5987.

27. Spiropoulou CF, Kunz S, Rollin PE, Campbell KP, Oldstone MB (2002) New Worldarenavirus clade C, but not clade A, B viruses, utilizes alpha-dystroglycan as its majorreceptor. J Virol 76:5140–5146.

28. Radoshitzky SR, et al. (2007) Transferrin receptor 1 is a cellular receptor for New Worldhaemorrhagic fever arenaviruses. Nature 446:92–96.

29. Martinez MG, Cordo SM, Candurra NA (2007) Characterization of Junin arenavirus cellentry. J Gen Virol 88:1776–1784.

30. Oldstone MB (2002) Arenaviruses II: The molecular pathogenesis of arenavirus infec-tions. Curr Top Microbiol Immunol 263:V–XII.

31. Daniels TR, Delgado T, Rodriguez JA, Helguera G, Penichet ML (2006) The transferrinreceptor part I: Biology and targeting with cytotoxic antibodies for the treatment ofcancer. Clin Immunol 121:144–158.

32. Gatter KC, Brown G, Trowbridge IS, Woolston RE, Mason DY (1983) Transferrin receptors inhuman tissues: their distribution and possible clinical relevance. J Clin Pathol 36:539–545.

33. Terrell TG, Stookey JL, Eddy GA, Kastello MD (1973) Pathology of Bolivian hemorrhagicfever in the rhesus monkey. Am J Pathol 73:477–494.

34. Jefferies WA, et al. (1984) Transferrin receptor on endothelium of brain capillaries.Nature 312:162–163.

35. Peters CJ, Zaki SR (2002) Role of the endothelium in viral hemorrhagic fevers. Crit CareMed 30:S268–S273.

36. Soda R, Tavassoli M (1984) Transendothelial transport (transcytosis) of iron-transferrincomplex in the bone marrow. J Ultrastruct Res 88:18–29.

37. Andrews BS, et al. (1978) Replication of dengue and junin viruses in cultured rabbit andhuman endothelial cells. Infect Immun 20:776–781.

38. Flanagan ML, et al. (2008) New World clade B arenaviruses can use transferrin receptor1 (TfR1)-dependent and independent entry pathways, and glycoproteins from humanpathogenic strains are associated with the use of TfR1. J Virol 82:938–948.

39. Oldenburg J, Reignier T, Flanagan ML, Hamilton GA, Cannon PM (2007) Differences intropism and pH dependence for glycoproteins from the Clade B1 arenaviruses: Impli-cations for receptor usage and pathogenicity. Virology 364:132–139.

40. Reignier T, et al. (2006) Receptor use by pathogenic arenaviruses. Virology 353:111–120.41. Rojek JM, Spiropoulou CF, Kunz S (2006) Characterization of the cellular receptors for

the South American hemorrhagic fever viruses Junin, Guanarito, and Machupo. Virol-ogy 349:476–491.

42. Wang E, Albritton L, Ross SR (2006) Identification of the segments of the mousetransferrin receptor 1 required for mouse mammary tumor virus infection. J Biol Chem281:10243–10249.

43. Lawrence CM, et al. (1999) Crystal structure of the ectodomain of human transferrinreceptor. Science 286:779–782.

44. Bennett MJ, Lebron JA, Bjorkman PJ (2000) Crystal structure of the hereditary haemo-chromatosis protein HFE complexed with transferrin receptor. Nature 403:46–53.

45. Cheng Y, Zak O, Aisen P, Harrison SC, Walz T (2004) Structure of the human transferrinreceptor-transferrin complex. Cell 116:565–576.

46. Kastello MD, Eddy GA, Kuehne RW (1976) A rhesus monkey model for the study ofBolivian hemorrhagic fever. J Infect Dis 133:57–62.

47. McKee KT, Jr., Mahlandt BG, Maiztegui JI, Eddy GA, Peters CJ (1985) ExperimentalArgentine hemorrhagic fever in rhesus macaques: Viral strain-dependent clinicalresponse. J Infect Dis 152:218–221.

48. Webb PA, Johnson KM, Mackenzie RB, Kuns ML (1967) Some characteristics of Machupovirus, causative agent of Bolivian hemorrhagic fever. Am J Trop Med Hyg 16:531–538.

49. Webb PA, Justines G, Johnson KM (1975) Infection of wild and laboratory animals withMachupo and Latino viruses. Bull WHO 52:493–499.

50. Tesh RB, Jahrling PB, Salas R, Shope RE (1994) Description of Guanarito virus (Arena-viridae: Arenavirus), the etiologic agent of Venezuelan hemorrhagic fever. Am J TropMed Hyg 50:452–459.

51. Warrens AN, Jones MD, Lechler RI (1997) Splicing by overlap extension by PCR usingasymmetric amplification: An improved technique for the generation of hybrid pro-teins of immunological interest. Gene 186:29–35.

52. Kenyon RH, et al. (1992) Aerosol infection of rhesus macaques with Junin virus.Intervirology 33:23–31.

Radoshitzky et al. PNAS � February 19, 2008 � vol. 105 � no. 7 � 2669

MIC

ROBI

OLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

020